And how MPC had to make Seth Rogen’s Uncanny Valley Bob character look as bad as possible.
Akiva Schaffer’s Chip ‘n Dale: Rescue Rangers, currently streaming on Disney+, is in many ways a love letter to animation, even though it also offers a tongue-in-cheek perspective on the hand-drawn versus CG side of animation, and on the role of the so-called Uncanny Valley in animation and visual effects.
Interestingly, the film–which re-teams the famous cartoon characters Chip and Dale (with Dale now having had ‘CGI surgery’)–was brought to life in many ways, including with classic 2D hand-drawn animation. But to get the film done on time and on budget, much of the animation for the principal characters, even if they appeared to be ‘2D’, was still realized in 3D, and then treated in rendering and comp with a toon-like shader.
This work was handled by MPC, which also brought to life CG characters such as Bob (Seth Rogen), a clear reference to the existence of Uncanny Valley motion capture characters from several years ago, with one challenge being having to degrade their normal photoreal efforts to an acceptable degree to sell the effect. The studio further crafted many entirely synthetic environments for the live-action film.
In this befores & afters interview, MPC visual effects supervisor Steve Preeg breaks down the key character work, filming stand-ins and stuffies for the shoot, and the challenges of mixing live-action and cartoon in a world where this interaction is all very ‘normal’.
b&a: What approach did you have, going into this film, about how the mix between 2D and 3D would be achieved?
Steve Preeg: The 2D was definitely the elephant in the room. If you look back at something like Roger Rabbit and the budget and time to make a movie like that, where you’re putting traditionally hand-drawn characters into the real world, it’s insanely expensive and time-consuming and very restrictive in terms of the editorial process. You have to lock the movie a year before the release.
Here, we were still making significant changes to the main characters two and a half months before we were meant to be done. And so, there was, instead, a lot of discussion about using 3D animated characters with a cell-shaded look to them. Not only because of the time constraints, but also budget. There are still over a hundred hand-drawn characters in the movie. We just knew that we couldn’t do the main characters this way.
So, we did a lot of investigating in terms of, how can we try to pay homage to the 2D animation side? You’ll see that from a perspective standpoint, things like the tufts of hair on Chip’s cheeks, they don’t turn, and moving through, they’re always a silhouette-y kind of thing. It was all hand-drawn in comp, putting lighting lines in, adding quickly hand-drawn gradients in. There was a lot of actual hand-drawn effort put on top of the 3D work.
b&a: So, the 2D principal characters that look 2D are actually 3D animated characters with a 2D shader on them?
Steve Preeg: Chip and Gadget and Dale before he had the surgery, and Monty and Zipper–those are all 3D assets that were done with a 2D shader. When I say that, ‘with a 2D shader,’ there was a lot of custom shot effort done. If you take a CG character that’s 2D and you either rotate them or rotate a camera around them, there’s a much more significant feeling of perspective change than what we were aiming for. So, we were doing a lot of tricks to try and make it feel that we were trying to at least adhere to the feeling of 2D. We know we’re not going to fool everybody, and we know there’s going to be the traditionalists who say, ‘Well, why didn’t you just hand-draw it?’ If we had another double the budget for visual effects and an extra year of post, maybe we could have.
b&a: Maybe ignorantly or naively, I thought perhaps that was an actual 2D animated approach where a different studio had come in and done them next to MPC’s work?
Steve Preeg: We did have a 2D team for the 100-plus other characters. That team was also doing draw-overs. They were doing style sheets and guides and character studies for the 3D animators. They did a bunch of targeted poses and targeted shapes for the 3D animators to work on.
b&a: When you work on any creature show, there’s storyboarding and previs, and then the shoot. What was the workflow that the filmmakers followed here in terms of visualization?
Steve Preeg: It was the same process. We had a full storyboard reel done. I think the film has maybe 40 shots without visual effects in it. So, boarding and previs’ing the entire film made sense. Akiva being both the director and with an acting background, he voiced all the characters himself for the most part in the story reel. So, he was able to cut the movie with voice, with the pacing and ideas that he had in his head. We then previs’ed the entire film.
Probably slightly earlier than you might on other projects, we brought on our DP and our editor, our picture editor, as early as we possibly could. Larry Fong, being the DP, could get involved and work with us on camera. Nearly half the movie was fully digital. Of course, Akiva wanted it to feel like you’re not watching a visual effects movie with two small characters. He wanted this to feel like Lethal Weapon or Bad Boys. So, the camera work and the feel of the movie were meant to feel like we had just shot this as though we are their size. They’re not these cartoons that are outcasts in the real world. They just fit into the world. There are small cars. There are small houses. So, in theory, there are small cameras.
It was meant to be sort of this ‘California Noir’. And so, if you watch the film, there’s quite a bit of close-up, dialogue-y, almost very Bad Boys-y kind of stuff. I think that Akiva did a great job of wanting it to feel like you’re just watching a movie with people in it versus watching a movie with little characters. Even the scene in Monty’s apartment when Putty is there, and Ellie shows up for the first time and there’s that jump scare where she shows up in the window–it’s like this homage to King Kong. The idea was we want the audience to forget that we’re in this small world, and we’re just with these characters in their world. And then, boom, you remember, ‘Oh, right, right. We’re small. We forgot that we’re small for a minute.’
b&a: Was there a virtual production side of this imagined at all?
Steve Preeg: Yes and no. What happened was there was the intention to have a watered-down version of virtual production. Our executive producer had come off of Lion King and we happened to have access to a lot of the equipment that they had. In our pre-production offices, we had set up a small virtual production stage where we could do motion capture and camera capture simultaneously. We didn’t have a full-size motion capture stage, but we didn’t necessarily need it because we weren’t motioned capturing stuff for final production. We were motion capturing to get a feel for everything because it was basically going to be hand-animated anyway. But we could potentially keep some of the camera work. So we set up this small stage, we brought on our DP, and I think our DP was on the show for one week when COVID hit.
Basically, they said, ‘Okay, everyone goes home now.’ I called Larry and I said, ‘Hey, you may not really be interested in this, but if you are, we could set you up with Maya. We could give you a tutorial.’ He was super interested. So, we set him up with a copy of Maya. We would send him over blocked scenes for the fully CG shots and let him play around in Maya, and I think he had fun with that.
And then, near the end of pre-production, we were allowed to go back in with very small teams of two to three people at most. We weren’t doing motion capture at that point. It had to be very limited in manpower, so we would basically prep scenes and bring them in there with the animation already in and then just let him do camera work with the virtual stage.
b&a: For the shoot, what worked, and what did you do in terms of stuffies and stand-ins?
Steve Preeg: For Chip and Dale specifically, we had one stuffy for Dale. And then, we made the stuffy for Chip, but since he was going to be 2D, I just decided to make him be a gray stuffy to represent a diffuse ball. We did the standard clean plates and plates with the stuffies.
Moving shots were my big concern. When you have a one-foot-tall character whose legs are only probably two-and-a-half inches long, a walk-and-talk is really slow. My concern was that even if you tell the camera operators, ‘Look, no, I mean, slow. Like, you’re moving s-l-o-w,’ there’s a chance they will still go way too fast, and then we were going to have a bunch of plates that we either had to re-time or reconstruct. If there are live-action people in the background, retiming it is a pain.
And in fact, we did a camera test initially, and I said, ‘Look, guys, really slow now. Really slow.’ Then with all the plates, we had to have the chipmunks running full speed to even stay with the camera. So, what we did was, we had the special effects team make these LED ribbon strips that they could lay down very quickly, and then we did a bunch of different walks, run and jog cycles of the different speeds that the chipmunks could do. Say they’re really in a rush on all fours. What if they’re just walking? All these things. And then, they had pre-programmed all these speeds into their LED strip so they could just say, ‘Walking speed. Running speed. Jogging speed.’ Whatever it needed to be.
Prior to the shooting, I had gone through every shot and called out what the speed was going to be. On set, for each setup, we could be like, ‘Okay, they’re going this fast. They’re going from here to here. Lay out the LED strip.’ That would give the dolly grip a speed to move at. If they were just panning, we also then had sticks that would show the head height. We talked about putting the LED strip high enough up that it would represent their head height, but they were like, ‘We don’t really need that. It’s going to take more time for set up.’ As long as we have the speed on the ground that we can see, and then we have the height of their heads, we could match it. In the end, the plates for Chip and Dale were actually quite useful in almost every case because that had been planned out pretty well.
For some of the other characters, we had some stand-ins, too. We had a Putty stuffy that was the right size for him, who was also somewhat posable, so we could throw him in. For Bjornson, who was the cheesemonger, he was obviously somewhat based on the Swedish Chef, and the Henson Company was nice enough to bring the Swedish Chef on set. Not to be puppeteered or used directly, but as a lighting reference and a standee. So we had literally had the Swedish Chef on set, which was pretty cool. For a lot of the other characters that were man-sized, we actually had people that we prepped out and put cartoons over them.
In fact, a little bit of a funny story: if you watch at the very end of the movie when Ellie says that she’s going to start her own detective agency, she hands the bag with Putty over to an FBI agent who’s like a jackal or some kind of dog cartoon character. That was originally going to be just a live-action person. And when we watched the final edit of the movie, Akiva realised that other than just people in the background, there was not any shot in the movie where a human being spoke to another human being. And he said, ‘I don’t want to ruin it at the end, so we need to turn that guy into a cartoon. Because I don’t want a human to ever talk to another human.’ They’re always interacting from a cartoon to a human. So we erased that guy’s head and put it in a cartoon.
b&a: What about for characters like Sweet Pete and Bob? Was there someone performing those on set?
Steve Preeg: We had a stand-in for Sweet Pete who was the correct height as Peter Pan, at five foot two. We had a small person to play Bob. We actually had a couple of stand-ins, because the guy who was more of an actor was not quite the right size. We also had a couple of guys playing Jimmy the polar bear, one who was really tall, and then we had another guy who was tall, and where we added a stick thing above his head to make him taller.
b&a: When it came time to do animation, I felt like sometimes the characters were on ones. Sometimes they’re on twos. It depended on what ‘character’ they were. It was very effective. But have I got that right?
Steve Preeg: Someone like Dale is always obviously on ones. He was just a CG character. For the Roger Rabbit cameo, we put him on ones because the original Roger Rabbit was on ones because they spent the money. Chip was supposed to be from the original show, so he was on twos. However, there were times when he’s interacting with something that is physically there, or there were times when you have moving cameras–and you don’t see a lot of moving cameras in 2D animation because that can get really stuttery–so there were times where we would have to put his global animation on ones, but then his internal movements of arms and head and stuff could be on twos.
When it came to other characters, like the cameo of the Transformer or the guy in the locker room in the Speed Stick ad was supposed to be a GI Joe-ish type era character. Those were not only on twos, but also drawn in a style that was similar to that era. His body may be on tens and his heads on twos, because they would do that back in the day to save money. It’d be like, ‘Alright, this guy’s just sitting there talking. So, his head can move, but you don’t have the subtle rotations of body.’
b&a: In terms of Chip and turning him from a 3D character into 2D, what did you find were some of the challenges for tone mapping or contrast or shadows or all the things that we associate with 2D? What did you find with the rendering, but also compositing, challenges there?
Steve Preeg: All of that becomes very challenging when you’re trying to make it adhere to the 2D world but fit into a 3D world. For example, having just the variation of lighting that you can have from a blown-out side of Dale’s fur to a shadow side that’s quite dark, say when they’re in the cheese cave. You only get a couple of values for Chip, right? You can have a key and you can have a shadow and then a mid-tone. How bright do you make that key? Because the brightest part on Dale is almost blown out, but that’s got to transition to something else. So where is that line that you put that in at? There were a couple of different tricks we used. One was using those terminator lines that give you a clear line, which they didn’t have in the Chip ‘n Dale TV series. It was too simple back then. And cost-wise, they wouldn’t do that. But we had to put something there, or else he just looked very flat.
When there were even larger lighting changes, just those terminator lines with a couple of value changes wouldn’t work. We would hand-draw in, in comp, slight gradients across the characters just so that within the mid-tone range, you would have a difference between this cheek and this other cheek. If your key ends here and your shadow starts here, it would give you just a slight difference across the face. But we didn’t want it to follow the contours of the body, because then it would feel too 3D. So, it was just a lot of hand-drawn stuff in comp. The key lines, the gradients, all that stuff was just added in 2D in comp.
b&a: One of the things that I want to ask you about is the Uncanny Valley side of this. Clearly, you’ve worked on projects where you’ve overcome that and had that as a challenge at Digital Domain and MPC. And here, there’s Bob and the whole reference to some of those older motion capture shows. I mean, it’s fascinating and it’s hilarious, and I love that it’s kind of treated like that. But just personally, but also technically, what kind of challenge was that in terms of not making the character too good?
Steve Preeg: That was actually a huge problem. Bob was the character that we did in a proof of concept for this, as well. And I’ve been through all zones of the Uncanny Valley in my career. Final Fantasy had plenty of stuff in there. TRON: Legacy had some scenes. Benjamin Button, though it was well-received, to me, still has plenty of scenes in that. I had just come off, with many other visual effects supervisors, the film Cats. I think Akiva threw that in there as a little jab at me.
So, when we first started doing Bob in the proof of concept, it was way too good. And Akiva kept saying, ‘Look, if you don’t make it bad enough, people will just think you’re bad at your job. They’ll think you tried to make it good, and you failed. So, it has to be bad enough that people know it was intentionally bad.’
We kept making baby steps towards making it bad. And he kept going, ‘Nope, not even close. Not even close. Not even close.’ He goes, ‘If you guys don’t make it bad, I’m going to use the previs character in the final movie. It needs to be a lot worse.’ It’s like drawing blood from a stone to get a VFX artist to just make something that sucky.
And then, it was a different crew that did the final movie versus the proof of concept. So then, we had to go through it all again. And again, it was the same thing, where it was like, ‘It needs to be worse. It needs to be worse.’ And, ‘I will use the previs!’
We had a screening. We had a last-minute cancellation and they said, ‘Hey, do you have a friend who would want to come for the screening?’ I had a friend who’s a voice actor in LA, and I called him up and I said, ‘Hey, what are you doing? You need to be at the Disney lot in 10 minutes.’ And he’s like, ‘Okay, I’m on my way.’ And he watched the movie. And at that point in the film, there were shots of Bob that were pretty much done, and there were shots of Bob that were still postvis.
There was a discussion at the end, as they often have at these things, and my friend said, ‘You know, I thought that the Bob character was much more successful when it was just the postvis.’ Now, it’s all of a sudden like, I have a lot more work to do…
But he was right. We just had to keep making him worse so that, at least hopefully, most of the audience realizes that it was intentional. I was really worried that he was in the first trailer because in my head it was a bit out of context. I was like, ‘Are people going to get it? I don’t know.’
b&a: Well, the dead eyes and not looking at the other characters helps as well, I guess.
Steve Preeg: Yes, but in the trailer, we weren’t allowed to say, ‘Polar Express.’
b&a: Right.
Steve Preeg: We weren’t allowed to reference that as a movie. And we didn’t mention the Uncanny Valley in the trailer either. So, there wasn’t quite as much context for it, but it seems like people got it for the most part. I think there were a handful of things that I saw where it was just like, ‘Man, some of the CG in the movie’s bad.’ And we were like, ‘Mm-hmm. Yeah, on purpose.’
b&a: Chip’s dog, is that real or CG?
Steve Preeg: It’s CG. It was based on our executive producer’s dog, so we scanned that dog and built him. We had talked about using something in MPC’s element library, but there are a lot of scenes in the movie that are fully CG, and pretty much anytime we are in a world where we live at their scale, like the interior of Chip’s house, for example. The cheese cave was all digital. Inside Monty’s apartment was all digital. It made sense to do the dog digitally.
b&a: You mentioned of course the hand-drawn animation approach would’ve taken more time. But was there ever any consideration that the characters that are stop-motion-looking would really be stop-motion? Did that ever come up as, ‘Let’s do Putty stop motion?’
Steve Preeg: I think it was briefly talked about. Again, the first time that you see Putty, he is in Monty’s apartment, which was never going to be shot. We had a very limited shoot schedule and stop-motion is extraordinarily time-consuming. And then there was the fight between him and Ellie at the end. Stop-motion interaction with a live-action human is–I’m not going to say impossible–because I’ve seen say The Secret Adventures of Tom Thumb, but in that, the people are moving stop-motion, as well.
At some point, Putty was going to have to be digital anyway. That fight went through a lot of iterations. It was very violent in the first round, but the feeling was, that we don’t really want to have Putty beating up a woman. So, it was like, ‘Okay, how do we make him still feel very threatening, but have this fight a lot more fun?’ We really opened it up to postvis to just go crazy.
They come up with a lot of that kind of interesting stuff. Like the sock puppet that showed up. That was just one of the previs guys who made a sock puppet and shot it on a green screen and put it in the shot, and everyone laughed and was like, ‘Okay, well, we’re keeping the sock puppet.’ And when Putty’s hands turn into cymbals and start smashing her, the first time I saw that I’m like, ‘Alright, who’s idea was this? This is insane.’ And then, Akiva saw it and he’s like, ‘That’s fun!’ So, it stayed in.
